Download Purpose Radiofrequency (RF) ablation is an effective

ATRIAL WALL THICKNESS IMAGING FOR CAVOTRICUSPID ISTHMUS ABLATION
Tobias Voigt1, Peter Koken2, James Harrison3, Steffen Weiss2, Sascha Krueger2, and Tobias Schaeffter3
Clinical Research Europe, Philips Research, London, London, United Kingdom, 2Tomographic Imaging Systems, Philips Research, Hamburg, Hamburg, Germany,
3
Division of Imaging Sciences, King's College London, London, London, United Kingdom
1
Purpose
Radiofrequency (RF) ablation is an effective treatment for patients with atrial fibrillation and atrial flutter showing less
recurrence of arrhythmia than antiarrhythmic medications. The success of the RF-ablation procedure depends on the
creation of transmural lesions that block unwanted conduction pathways. The RF-ablation power has to be adjusted to
create transmural lesions, while avoiding perforation of the heart. This is challenging since the patient’s individual local
atrial wall thickness is unknown. MR wall thickness imaging (WTI) may provide this information and contribute to
increased success rates and safety [1]. In this work a fully integrated WTI procedure as input for cavotricuspid isthmus
(CTI) ablation is described. The technique is validated in a phantom study and applied in healthy volunteers and an atrial
flutter patient.
Methods
WTI was performed in five steps. 1) A 3D whole heart scan (400×280×165 mm3, resolution 2×2×3 mm³, acquisition
window: 130 ms, TE / TR = 2.7 / 5.3 ms, ECG triggered at mid diastole) was acquired (Fig. 1A). 2) Based on the 3D scan,
a model-based, automatic segmentation of the different heart chambers was performed [2] (Fig. 1B). 3) The two spatial
endpoints of CTI (tricuspid valve (TV) and inferior vena cava (IVC)) were selected manually on the segmentation surface
of the right atrium. A corresponding scan geometry including CTI and read-out direction perpendicular to the atrial wall
was generated without user interaction (Fig. 1C). 4) In this scan geometry, a 2D black-blood single shot turbo spin echo
(TSE), fat suppressed zoom image, with perpendicular slice selection and refocusing gradients [3], which restrict the FOV
in two dimensions was applied (FOV=300×120×3mm³, voxel size = 0.5×0.8×3mm³, 98ms acquisition window, TEeff = 4.9
ms, ECG triggered at mid diastole, Fig. 1D). 5) Wall thickness was defined as the FWHM of the central signal intensity
profile peak. Wall thickness was calculated for all positions along the CTI.
The accuracy of WTI imaging was tested in a resolution phantom with defined wall sizes (1mm – 18mm, Fig. 1E). The
effect of partial volume on WTI was investigated by diverting the imaging angle from the optimal perpendicular position
(0-50°). The WTI procedure was performed in five healthy volunteers and one atrial flutter patient. All experiments were
performed with a 32-channel cardiac coil on a 1.5T scanner (Philips, Best, The Netherlands) connected to the
Interventional MRI Suite (Philips Research, Hamburg, Germany) to implement the WTI workflow.
Results & Conclusion
The phantom study revealed, that WTI can reach submillimetre accuracy if planned perpendicular to the atrial wall (<10°
deviation, Fig. 2). A representative wall thickness profile along CTI is shown in Fig. 3. The local wall thickness along CTI
varied from on average 1.9mm near the TV to 5mm near the IVC. In future studies the study population should be
increased to investigate how atrial wall thickness changes between individuals.
References: [1] Koken, Proc.ISMRM 2011, p3734. [2] Heese, Proc.ISMRM 2009, p4681 [3] Stehning C, Proc.ISMRM 2010, p1409
Fig 1 (top): Wall thickness imaging process as
proposed in this study. A) 3D whole heart scan. B)
Automatic segmentation of the heart (based on A).
CTI endpoints are selected manually on the surface
C) Perpendicular scan geometry on the surface of the
right atrium (RA). D) Zoom image of CTI showing
aorta (AO), left atrium (LA), IVC and RA. E)
Phantom and zoom image from validation study.
Fig 2 (right): Errors in WTI due to partial volume
effects measured in a phantom. WTI performed
perpendicular to the atrial wall (0°) as in this study
can achieve submillimetre accuracy.
Proc. Intl. Soc. Mag. Reson. Med. 21 (2013)
Fig 3: Wall thickness profile measured
along CTI, starting from tricuspid
valve. A moving average is shown.
0470.